Cognitive State Classification using Optimized Feature Selection Approach

Functional Magnetic Resonance Imaging (fMRI) is the most widely used technique where it is possible to capture neural activity in human brain regions when subjected to different stimuli. However, due to the fMRI dataset’s high dimensional and sparse nature, the selection of appropriate features plays a crucial role in achieving the best classification accuracy. In this work, the stable features are selected from the fMRI dataset by combining Fast Fourier Transform (FFT) with Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). Then, the machine learning classifiers such as Support Vector Machine (SVM), Gaussian NB, and XGboost have been trained using these features. StarPlus fMRI dataset is used to examine the performance of the proposed feature selection framework. The experimental results reveal that the proposed feature selection algorithm resulted in optimum features with better classification accuracy. Comparison of the proposed scheme with state of the art models show that it performs better, and as a result, can be used for the pattern recognition of brain responses in multisubject fMRI data.